Comparison of precipitation measurements by OTT Parsivel2 and Thies LPM optical disdrometers

Event E365

1 Introduction and data set

Event	Start	End	Length	Mean I	Max I	Mean Z	Mean KE	Cum P	NP	D50
365 p1	25/11/2014 23:25	26/11/2014 03:04	220	11.31	240.51	28.4	10.6	40.6	415941	0.444
365 p2	25/11/2014 23:25	26/11/2014 03:04	220	11.32	198.90	31.8	13.5	39.6	460871	0.448
365 p3	25/11/2014 23:25	26/11/2014 03:04	220	6.24	111.00	27.1	13.1	21.8	83125	0.856
365 p4	25/11/2014 23:25	26/11/2014 03:04	220	5.70	96.43	27.1	12.7	21.0	72017	0.880

  # define start and finish times
  a <-as.POSIXct(strptime("2014-11-25 23:30:00","%Y-%m-%d %H:%M:%S"))
  b <-as.POSIXct(strptime("2014-11-26 04:00:00","%Y-%m-%d %H:%M:%S"))
  times <- seq(a, b, by="min")

Read the raw PSVD data, using function psvd_read from the disdRo package.

# Thies 1
files <- list.files('./p1','.txt$|.txt.gz$', full.names=TRUE, recursive=TRUE)
dsd1 <- psvd_read(files, type='Thies')
dsd1 <- apply(dsd1, c(2,3), sum)

# Thies 2
files <- list.files('./p2','.txt$|.txt.gz$', full.names=TRUE, recursive=TRUE)
dsd2 <- psvd_read(files, type='Thies')
dsd2 <- apply(dsd2, c(2,3), sum)

# Parsivel 1
files <- list.files('./p3','.txt$|.txt.gz$', full.names=TRUE, recursive=TRUE)
dsd3 <- psvd_read(files, type='Parsivel')
dsd3 <- apply(dsd3, c(2,3), sum)

# Parsivel 2
files <- list.files('./p4','.txt$|.txt.gz$', full.names=TRUE, recursive=TRUE)
dsd4 <- psvd_read(files, type='Parsivel')
dsd4 <- apply(dsd4, c(2,3), sum)

---

2 PSVD plots

Particle size and velocity plots, using function psvd_plot from the disdRo package. First, we shall plot the raw data as read from the disdrometer telegrams, and then we shall apply a filter to the PSVD matrix based on a 50% difference with respect to the fall velocity model of Beard (1976).

2.1 Unfiltered data

2.2 Filtered data

We first create a filter using the function psvd_filter from the disdRo package. We limit the data to particle sizes between 0.25 and 8 mm. We use a discrepancy threshold of 50% with respect to the Beard (1976) model, setting the elevation (alt) to 230 m to account for the elevation effect on the fall velocity of raindrops.

There will be two filters, one for Thies disdrometers (fltT) and one for Parsivel ones (fltP).

# Thies
fltT <- psvd_filter(type='Thies', d=c(0.25,8), tau=0.5, alt=230)
image(fltT)

# Parsivel
fltP <- psvd_filter(type='Parsivel', d=c(0.25,8), tau=0.5, alt=230)
image(fltP)

We now do the plots. Transparency of the filtered area can be set with parameter alpha.

3 Particle size distribution

We now produce particle size distribution plots, using the function psd_plot from the disdRo package. These plots depict the particle density (ND), against the particle size class.

3.1 Unfiltered data

3.2 Filtered data

4 Particle velocity distribution

We now produce particle fall velocity distribution plots, using the function psd_plot from the disdRo package. These plots depict the particle density (ND), against the particle size class.

4.1 Unfiltered data

4.2 Filtered data

---

5 Event time series

5.1 Time series, measured variables

5.2 Time series, calculated variables, no filter & no margin correction

5.3 Time series, calculated variables, filter & margin correction